Sat Aug 25 10:42:23 CEST 2007
in fact, i need to distinguish between syntax and semantics a bit
better. a compiler works on syntax, (a representation).
von thun has some text about this..
again, i'm amazed by how untyped you can be in scheme! i'm just
performing operations on lists, without ever having to clarify what
things are.. interpretation is a consequence of what functions you
apply on the symbols..
so, let's say that "working with symbols" is always untyped. they are
a universal tool of delayed semantics. maybe that's the idea behind
formal logic, right? by just specifiying HOW to operate on symbols,
you never need to explain what you are actually doing.
Quite an adventure, trying to provide a model for the language and
* read Flatt's paper about macros again
* logic and lambda calculus.
* monads and their relationship with compositional programs.
* a purrr module system + compositional language
Funny. I have that book on my shelf, and i tried to start reading it
on thursday. I guess it has a major truth. Once the necessary
structure is in place, the conclusions are often trivial. So all the
effort is in the creation of structure. Sounds like programming.
Try "Once things are clearly defined, the solution is at most a single
line.", "Write the language, and formulate your solution in it.", "Ask
the right question."